How do we combat the misuse of AI and its effects on the internet and society?

[Janky]

I've been thinking about this for days now, maybe even weeks, months? I've become tired of this constant game of sorting the simulated from the simian, searching images and videos for imperfections that reveal their inhuman origins.

What can be done to stop this? It's miserable, watching the sites I grew up on be slowly consumed by this wave of fake people, and it's not like we can just hide here and ignore the problem, it'll just creep into here as well!

What do you, fellow (hopefully real) people of this beautiful forum I call home, believe can be done to stop this catastrophe? Do we keep descending deeper into the depths of the hyper-curated internet?

Personally, I don't think the solution to the AI takeover comes from heavier and heavier regulations, I think that a large part of AI's dominance comes from that over-regulation (coming from an econ nerd) and that those regulations created the market incentives for large tech companies to invest incomprehensible amounts of money into a technology that might otherwise not be so profitable.

Still, I want to hear from y'all! I want to have a civil debate about this! It's important we have these conversations while there's still time to have them!

-Janky

[elsieee]

I'm probably being overly pessimistic but, honestly, I think it's inevitable that AI will eventually become ingrained into a part of the internet lifestyle (bar small forums like these maybe).

Tech companies are now more pushy than ever to put AI in everything. Google search, social media, emails, and every other function that we can't avoid. It's clear that they want AI to be a part of the new norm, and insofar as these tech giants keep pushing for it, there's nothing much we can do to really have an impact on the AI issue.

And it's not just these tech giants, lots of people are behind AI too (just think about the Ghibli art style and the 'starter pack' trends) because it's something that's almost magical to them, being able to answer all their questions and create every image. Whenever I ask my older family members a question they don't know, they'll most likely say, "let's see what ChatGPT says." Sure, in a sense, generative AI makes knowledge more accessible, but you can see how scary it is that people are increasingly reliant on it for knowledge, among other things.

So, insofar as most people still think AI is good, or even if it's just a 'useful tool,' you can't really stop the effects of it. Those tech giants and governments will only consider a voice if it's the voice of the vast majority, and with AI it's clearly not happening anytime soon. And the later it's not dealt with, the less likely it will be.

One can only wish that, sometimes in the near future, the voices against AI can be picked up by mainstream parties and media too, instead of just being voices echoed on the internet. Until then, we can only really protect our own work and our own websites from AI.

I'm curious to hear what others think of this issue as well 

[Janky]

I'm going on a bit of an economic/political tangent for this one, please so bear with me and please, please keep an open mind. This is coming from a place of honesty and care, please don't take this as me being political for the sake of being political, policy is important dang it and I'm too big of a nerd not to talk about it! Please be kind!

So, some background. I'm in college and I'm going into a major for economics (ik, exciting). In class we discussed many things, and in macro-economics specifically we discussed a lot of government policy and how it affects the economy, how markets only fail when governments get involved in some way and meddle with the market, etc etc. It would be an understatement to say that this class was impactful to me, it completely changed how I saw the world and the way people live in it.

Keep this in mind, I am a supporter of free market policies. I work on economic logic. I got curious about how a decentralized world would work, if you stripped away the central government. Typical FMC stuff, most of it is not important to my point.

Here is where AI and freedom of information becomes important.

I thought to myself, how would intellectual property work in a FMC society? I searched for a way, tried devising ways of it working, but all of them involved a central authority. Eventually I caved and asked my econ professor, and his reply made me realize something that most people don't enjoy me talking about. Please don't hate me for this. I really hope I don't get banned for this post.

His answer was simple. Basically, because information (that includes ANY digital content) has no scarcity, or in layman's terms it can be copied infinitely without physically stealing it from another person, it cannot be "owned" in the way that we think about property ownership. It can be created and destroyed, but never owned like one might own a car or a house or a sandwich. The argument is of course way more nuanced than that, but most of it is not important here. The important part is that intellectual property cannot exist in a free market society, and that abolishing it would do away with a system that does more harm than good.

It is this, this singular axiom, that could solve the AI crisis. The reason that corporations have any incentive to develop these AI's is because they think they can make profit off of them. They have a monopoly over information thanks to their copyrights and patents on their AI's and their architectures, and thus they can make money off of them with little competition. If this is changed, if intellectual property is abolished, it has the side effect of removing that incentive. The profitability of AI is already being brought into question with how expensive it is to run datacenters, and this would be the final nail in the coffin. Now, it would not completely remove AI as a threat, but it would greatly diminish it. There would be no reason for companies to maintain enormous data-centers to run complex models if anyone could copy the code and run a simpler version locally.

However, the removal of intellectual property would also have an effect on the internet as a whole, as massive social media giants would crumble without the ability to maintain protection of their advanced algorithms or make money from advertising or paid subscriptions. This has the potential to decentralize the internet, which in turn would make it much harder for AI's to take it over as the average internet space becomes smaller, more easily manageable, and more curated for humans. Think about MelonLand for instance, this space is resistant to AI's because it is smaller, more curated, more gated. It is the embodiment of the decentralized internet, and it serves as a prime example of how to create an environment where people can communicate without the dangers of AI.

Thank you for reading this utter ramble, and for letting me talk about this issue in some depth. If there is any silver lining to the takeover of AI, it might force people back into these small, curated forums, where AI cannot find them as easily.

Unrelated, I can't seem to figure out art submissions? Whenever I submit the images it just drags me back to the start where the boxes to put tags in are.

[elsieee]

The important part is that intellectual property cannot exist in a free market society, and that abolishing it would do away with a system that does more harm than good.

I think the solution you're talking about just boils down to how much you are willing to risk to solve the 'AI' crisis.

I'm going to assume that you're saying we should abolish the concept of intellectual property as a whole, and not just for internet stuff. In that case, I think the effect of doing so is going to be so immense that, as dangerous as the AI crisis is, is simply not worth it.

Intellectual property doesn't just apply to AI or media giants protecting and making money off of their algorithms and stuff, but also the devices that we use, the books that we read, the water bottle I have on my table right now (and just about everything else on that table), and basically nearly every artificial things in the world is protected by intellectual property laws in one way or another. Heck, if you look at the footer of the page, this forum is copyrighted too!

So I don't think we can simply 'do away' with intellectual property because it's the backbone of every creation and every innovation that we rely on in our daily lives, whether we like it or not. Because the tech giants aren't going to be the first ones to get impacted by the lack of protection, but the smaller creators. They are going to be the first ones to be discouraged from making things. And even if we don't care about them, the reason why innovation is possible is partially because of intellectual property. People are incentivised to make stuff better because they know their work won't be stolen without them at least making a profit. Vice versa, if today everyone can just take everyone else's creations without some legal mechanisms to protect said creations, then no one will wanna make stuff because the cost far outweighs the benefit.

so while abolishing intellectual property is a good solution to nip the ai problem in the bud, it's really only plausible in a vacuum

p.s. sorry if this post sounds critical I'm saying all this in a friendly debate way I swear :,)

p.p.s. there seems to be some art submission issue for a long time, you should let the mods know in the bot report forum!

[Grimlock]

I don't think the solution to the AI takeover comes from heavier and heavier regulations

personally i have to agree, even if training AI on other's works is legally made copyright infringement, companies like disney, adobe, etc will just make their own AIs with art, stock images, etc that they already own. so when it comes to people like freelance artists, who often already hand over rights to what they make over to companies, they would likely be paid to make a few things and then since the company would own whatever the final products were, they could let their AI train on it and make more than the freelance artist did for theoretical profit. while i believe AI works aren't themselves copyrightable, companies will have no reason to actually admit that they are and will likely copyright them anyway, especially when hiring people to clean up AI is possible. i believe a theoretical disney AI would probably replace concept artists before anything else, if its used later in the process, it would likely create a strange ship of theseus situation where one will have to ask, "at what point do AI works cease being AI works legally, once touched up by a human" in a court of law.

i do also have to agree that there may be no getting rid of AI for good, with or without intellectual property and copyright affecting it. if one strengthens copyright law by declaring AI infringement, should we trust corporations to enforce that fairly, especially as AI gets more convincing? what will likely happen is that the onus of proof is thrust upon artists to prove they didn't make their works with AI, especially those who "style match" the original work. without it, it would run wild, but there would be no reasonable way to be fully rid of it anywhere either way, since AI checks are fallible, and they would get better and harder to notice with time.

Also even with intellectual property and copyright existing, in order to actually defend what you own, you will likely have to get a lawyer, whether that be your own or someone else's to properly enforce it. While a DMCA is possible without one, some website's DMCA protections may be flimsy, if somehow a valid counternotice gets filed, you are out of luck if you can't get a lawsuit going in time. not even mentioning that you have to be aware that that your rights have been violated in the first place, since AI doesn't actually store 13 billion training items, one would have to prove that your work was actually in the dataset in the first place, made harder by the fact that the AI companies will likely cease storing the original images/text/etc as soon as they don't need them anymore. and if one is only talking about the US, its jurisdiction will only extend so far and won't protect one from everything. While sites like archiveofourown may defend your written work on its platform, folks in web revival circles who would like to host their works on their own sites, self hosted or not, will continue to be at risk. lawyers cost a lot of money to hire, money that many artists just simply won't have.

sorry this was mainly about corporations, i personally find more of the issue with AI lies with them and less individual people who might use an AI personally. Some AIs only take about as much electricity now as your average gaming session, which has thrown the power consumption angle out the window for individual uses. although i still wouldn't call it "good" as "individual uses" can vary wildly. i think AI is a really rough situation to handle, because corporations will take a mile if we allow it.